To inform the design of haptic information displays for noisy environments, we investigated two mechanisms for temporal masking of vibrotactile stimuli (backwards and common-onset) using a commodity display. We used a two-channel setup, presenting stimuli to the middle and ring finger of a participant's right hand. The stimuli consisted of 250 Hz sinusoidal waveforms displayed at a fixed amplitude in various combinations of duration (0, 30 or 300 ms) and stimulus onset asynchrony (0 or 30 ms). In anticipation of future embedded applications where signals are deliberately masked but levels cannot be individualized, signals were standardized at conservative (harder to mask) levels. Our results confirm the existence of a statistically significant masking effect for both forms of haptic masking explored, with common-onset exhibiting a significantly larger masking effect than backwards. However, an analysis of confidence in response levels shows no difference between the two successful masking techniques. We discuss mechanisms that could be responsible for these results, which have implications for the design of user interfaces that rely on tactile transmission of information.